Re: [f2fs-dev] [PATCH v5] f2fs: add reserved blocks for root user

[Chao Yu]

On 2018/1/3 10:21, Jaegeuk Kim wrote:
> This patch allows root to reserve some blocks via mount option.
> 
> "-o reserve_root=N" means N x 4KB-sized blocks for root only.
> 
> Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
> ---
> 
> Change log from v4:
>  - fix f_bfree in statfs

Could you fix f_bfree calculation issue in another patch prior to this
patch? That will be better for history tracking of patches or git bisect
when backtracking issues.

One more thing, should we move reserve_root_limit check to parse_option?
now, it looks that during remount we can set root_reserved_blocks exceeding
our defined limitation.

Thanks,

> 
>  fs/f2fs/f2fs.h  | 26 ++++++++++++++++++++++----
>  fs/f2fs/super.c | 34 +++++++++++++++++++++++++++++-----
>  fs/f2fs/sysfs.c |  3 ++-
>  3 files changed, 53 insertions(+), 10 deletions(-)
> 
> diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
> index 07e03990420b..a0e8eec23125 100644
> --- a/fs/f2fs/f2fs.h
> +++ b/fs/f2fs/f2fs.h
> @@ -95,6 +95,7 @@ extern char *fault_name[FAULT_MAX];
>  #define F2FS_MOUNT_PRJQUOTA		0x00200000
>  #define F2FS_MOUNT_QUOTA		0x00400000
>  #define F2FS_MOUNT_INLINE_XATTR_SIZE	0x00800000
> +#define F2FS_MOUNT_RESERVE_ROOT		0x01000000
>  
>  #define clear_opt(sbi, option)	((sbi)->mount_opt.opt &= ~F2FS_MOUNT_##option)
>  #define set_opt(sbi, option)	((sbi)->mount_opt.opt |= F2FS_MOUNT_##option)
> @@ -1105,6 +1106,7 @@ struct f2fs_sb_info {
>  	block_t last_valid_block_count;		/* for recovery */
>  	block_t reserved_blocks;		/* configurable reserved blocks */
>  	block_t current_reserved_blocks;	/* current reserved blocks */
> +	block_t root_reserved_blocks;		/* root reserved blocks */
>  
>  	unsigned int nquota_files;		/* # of quota sysfile */
>  
> @@ -1554,6 +1556,12 @@ static inline bool f2fs_has_xattr_block(unsigned int ofs)
>  	return ofs == XATTR_NODE_OFFSET;
>  }
>  
> +static inline block_t reserve_root_limit(struct f2fs_sb_info *sbi)
> +{
> +	/* limit is 0.2% */
> +	return (sbi->user_block_count << 1) / 1000;
> +}
> +
>  static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool);
>  static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
>  				 struct inode *inode, blkcnt_t *count)
> @@ -1583,11 +1591,17 @@ static inline int inc_valid_block_count(struct f2fs_sb_info *sbi,
>  	sbi->total_valid_block_count += (block_t)(*count);
>  	avail_user_block_count = sbi->user_block_count -
>  					sbi->current_reserved_blocks;
> +
> +	if (!(test_opt(sbi, RESERVE_ROOT) && capable(CAP_SYS_RESOURCE)))
> +		avail_user_block_count -= sbi->root_reserved_blocks;
> +
>  	if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) {
>  		diff = sbi->total_valid_block_count - avail_user_block_count;
> +		if (diff > *count)
> +			diff = *count;
>  		*count -= diff;
>  		release = diff;
> -		sbi->total_valid_block_count = avail_user_block_count;
> +		sbi->total_valid_block_count -= diff;
>  		if (!*count) {
>  			spin_unlock(&sbi->stat_lock);
>  			percpu_counter_sub(&sbi->alloc_valid_block_count, diff);
> @@ -1776,9 +1790,13 @@ static inline int inc_valid_node_count(struct f2fs_sb_info *sbi,
>  
>  	spin_lock(&sbi->stat_lock);
>  
> -	valid_block_count = sbi->total_valid_block_count + 1;
> -	if (unlikely(valid_block_count + sbi->current_reserved_blocks >
> -						sbi->user_block_count)) {
> +	valid_block_count = sbi->total_valid_block_count +
> +					sbi->current_reserved_blocks + 1;
> +
> +	if (!(test_opt(sbi, RESERVE_ROOT) && capable(CAP_SYS_RESOURCE)))
> +		valid_block_count += sbi->root_reserved_blocks;
> +
> +	if (unlikely(valid_block_count > sbi->user_block_count)) {
>  		spin_unlock(&sbi->stat_lock);
>  		goto enospc;
>  	}
> diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
> index 5c6a02b558f0..e814340bc2f0 100644
> --- a/fs/f2fs/super.c
> +++ b/fs/f2fs/super.c
> @@ -107,6 +107,7 @@ enum {
>  	Opt_noextent_cache,
>  	Opt_noinline_data,
>  	Opt_data_flush,
> +	Opt_reserve_root,
>  	Opt_mode,
>  	Opt_io_size_bits,
>  	Opt_fault_injection,
> @@ -157,6 +158,7 @@ static match_table_t f2fs_tokens = {
>  	{Opt_noextent_cache, "noextent_cache"},
>  	{Opt_noinline_data, "noinline_data"},
>  	{Opt_data_flush, "data_flush"},
> +	{Opt_reserve_root, "reserve_root=%u"},
>  	{Opt_mode, "mode=%s"},
>  	{Opt_io_size_bits, "io_bits=%u"},
>  	{Opt_fault_injection, "fault_injection=%u"},
> @@ -488,6 +490,18 @@ static int parse_options(struct super_block *sb, char *options)
>  		case Opt_data_flush:
>  			set_opt(sbi, DATA_FLUSH);
>  			break;
> +		case Opt_reserve_root:
> +			if (args->from && match_int(args, &arg))
> +				return -EINVAL;
> +			if (test_opt(sbi, RESERVE_ROOT)) {
> +				f2fs_msg(sb, KERN_INFO,
> +					"Preserve previous reserve_root=%u",
> +					sbi->root_reserved_blocks);
> +			} else {
> +				sbi->root_reserved_blocks = arg;
> +				set_opt(sbi, RESERVE_ROOT);
> +			}
> +			break;
>  		case Opt_mode:
>  			name = match_strdup(&args[0]);
>  
> @@ -994,21 +1008,21 @@ static int f2fs_statfs(struct dentry *dentry, struct kstatfs *buf)
>  	struct super_block *sb = dentry->d_sb;
>  	struct f2fs_sb_info *sbi = F2FS_SB(sb);
>  	u64 id = huge_encode_dev(sb->s_bdev->bd_dev);
> -	block_t total_count, user_block_count, start_count, ovp_count;
> +	block_t total_count, user_block_count, start_count;
>  	u64 avail_node_count;
>  
>  	total_count = le64_to_cpu(sbi->raw_super->block_count);
>  	user_block_count = sbi->user_block_count;
>  	start_count = le32_to_cpu(sbi->raw_super->segment0_blkaddr);
> -	ovp_count = SM_I(sbi)->ovp_segments << sbi->log_blocks_per_seg;
>  	buf->f_type = F2FS_SUPER_MAGIC;
>  	buf->f_bsize = sbi->blocksize;
>  
>  	buf->f_blocks = total_count - start_count;
> -	buf->f_bfree = user_block_count - valid_user_blocks(sbi) + ovp_count;
> -	buf->f_bavail = user_block_count - valid_user_blocks(sbi) -
> +	buf->f_bfree = user_block_count - valid_user_blocks(sbi) -
>  						sbi->current_reserved_blocks;
> -
> +	buf->f_bavail = user_block_count - valid_user_blocks(sbi) -
> +						sbi->current_reserved_blocks -
> +						sbi->root_reserved_blocks;
>  	avail_node_count = sbi->total_node_count - sbi->nquota_files -
>  						F2FS_RESERVED_NODE_NUM;
>  
> @@ -1136,6 +1150,9 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root)
>  	else if (test_opt(sbi, LFS))
>  		seq_puts(seq, "lfs");
>  	seq_printf(seq, ",active_logs=%u", sbi->active_logs);
> +	if (test_opt(sbi, RESERVE_ROOT))
> +		seq_printf(seq, ",reserve_root=%u",
> +				sbi->root_reserved_blocks);
>  	if (F2FS_IO_SIZE_BITS(sbi))
>  		seq_printf(seq, ",io_size=%uKB", F2FS_IO_SIZE_KB(sbi));
>  #ifdef CONFIG_F2FS_FAULT_INJECTION
> @@ -2571,6 +2588,13 @@ static int f2fs_fill_super(struct super_block *sb, void *data, int silent)
>  	sbi->reserved_blocks = 0;
>  	sbi->current_reserved_blocks = 0;
>  
> +	if (test_opt(sbi, RESERVE_ROOT) &&
> +		sbi->root_reserved_blocks > reserve_root_limit(sbi)) {
> +		sbi->root_reserved_blocks = reserve_root_limit(sbi);
> +		f2fs_msg(sb, KERN_INFO,
> +			"Reduce reserved blocks for root = %u\n",
> +				sbi->root_reserved_blocks);
> +	}
>  	for (i = 0; i < NR_INODE_TYPE; i++) {
>  		INIT_LIST_HEAD(&sbi->inode_list[i]);
>  		spin_lock_init(&sbi->inode_lock[i]);
> diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c
> index 93c3364250dd..ab6028c332aa 100644
> --- a/fs/f2fs/sysfs.c
> +++ b/fs/f2fs/sysfs.c
> @@ -162,7 +162,8 @@ static ssize_t f2fs_sbi_store(struct f2fs_attr *a,
>  #endif
>  	if (a->struct_type == RESERVED_BLOCKS) {
>  		spin_lock(&sbi->stat_lock);
> -		if (t > (unsigned long)sbi->user_block_count) {
> +		if (t > (unsigned long)(sbi->user_block_count -
> +					sbi->root_reserved_blocks)) {
>  			spin_unlock(&sbi->stat_lock);
>  			return -EINVAL;
>  		}
>